DFT Engine
Density functional theory calculations at B3LYP/6-311G++ and beyond. Parallelized across cloud GPU clusters with automatic queue management.
DFT Engine →
Quantum Chemistry Platform · Cambridge, MA
Screen 500 catalyst candidates before you book the fume hood
Qchemvyx runs DFT-accurate reaction pathway scans in hours — not weeks. Built for computational chemists who need throughput at the quality bar of B3LYP/6-311G.
The Problem
Traditional catalyst screening is a bottleneck in every R&D timeline
Synthesis, characterization, and fume-hood time are expensive. A single DFT scan on a local workstation takes 3–6 days per candidate. With 200–500 candidates in a design space, teams are forced to make arbitrary decisions about which structures to synthesize — not evidence-based ones.
The result: 80% of synthesis effort goes to candidates that could have been ruled out computationally within hours. Qchemvyx changes that equation.
Workflow comparison
| Metric | Traditional | Qchemvyx |
|---|---|---|
| Candidates/week | 5–8 | 500+ |
| DFT scan time | 3–6 days | ~8 hours |
| Fume-hood cost | ~$12K/week | Deferred |
| Throughput gain | — | 40× |
The Platform
Every layer of the quantum chemistry stack
Reaction Pathway Scans
Automated transition state search and IRC calculation. Full energy profiles for elementary reaction steps with NEB and QST3 methods.
Reaction Pathways →Screening Workflows
Define a candidate structure space, set DFT parameters, submit 500-structure batch — results ranked by activation energy.
Screening Workflows →Energy Landscape Analysis
3D potential energy surface visualization. Saddle point identification. Export to standard quantum chemistry formats including XYZ, SDF, and JSON.
Energy Analysis →Integrations
SMILES input, SDF/MOL2/XYZ. Output compatible with ORCA, Gaussian, VASP, and common ELN platforms including Benchling and LabArchives.
Integrations →Research-Grade Accuracy
Validated against NIST CCCBDB reference data and published CCSD(T) benchmarks. Not ML-approximated — actual DFT at the B3LYP/6-311G++ level.
View Benchmarks →
How It Works
From candidate library to ranked hits in hours
01
Upload Your Candidate Library
Submit SMILES strings or SDF files. Supported: up to 500 structures per batch. Automatic 3D structure generation with MMFF94 force field pre-optimization.
02
Configure DFT Parameters
Choose functional (B3LYP, ωB97X-D, PBE0) and basis set. Set solvent model, spin state, and convergence criteria. Saved parameter sets for repeat workflows.
03
Receive Ranked Energy Profiles
Activation energies, reaction energies, Gibbs free energy corrections. CSV export, interactive visualizer, direct API access. Results in ~8 hours.
Benchmark Validation
DFT accuracy benchmarked against CCSD(T) reference data
We validate every functional against coupled-cluster quality reference energies from 18 reaction classes in the GMTKN55 benchmark set. MAE values below 1.0 kcal/mol for thermochemical reaction energies.
View Full Benchmark Report →
GMTKN55 · Selected MAE values (kcal/mol)
| Reaction Class | B3LYP-D4 | ωB97X-D |
|---|---|---|
| Atomization energies | 0.64 | 0.51 |
| Reaction energies | 0.81 | 0.72 |
| Barrier heights | 1.42 | 0.89 |
| Noncovalent interactions | 0.55 | 0.48 |
Use Cases
Where Qchemvyx accelerates R&D
Heterogeneous Catalyst Design
Screen transition metal complexes and surface binding energies before synthesis. Rank by activation energy, selectivity, and thermodynamic stability.
Catalyst design workflows →
Process Chemistry Optimization
Identify rate-limiting steps and byproduct pathways in manufacturing reaction conditions. IRC scan for mechanism elucidation.
Process optimization workflows →
Materials Discovery
High-throughput electronic structure calculations for functional materials screening. Battery electrolytes, polymer additives, solid-state materials.
Materials screening workflows →
Early Research Partners
Used by computational chemists at research-intensive organizations
"We ran 400 amide coupling variants in a single weekend. That would have been 6 weeks of queue time on our local HPC."
"The activation energy rankings let us down-select from 320 candidates to 12 synthesis targets. The top-ranked hit turned out to be the best performer in the lab."
"What sets Qchemvyx apart is the transition state search quality. Not a heuristic approximation — actual NEB-driven saddle point location."
Ready to run your first 500-structure screen?
Request access. Our team will onboard you with a sample catalyst library within 48 hours.